future idea - WAV distortion analysis
Posted: Tue Feb 15, 2005 3:26 pm
I have an idea for a possible future GoldWave function - calculating the distortion of the WAV.
I noticed that by storing the edited WAV internally in 32-bit floating point format, peaks that exceed the 0 dB threshold (due to volume changes, etc.) are not truncated to 0 dB until you actually save the WAV file to disk.
I do a lot of level equalization from WAVs ripped from CDs, and many older CDs have a low overall volume with some peaks close to the 0 dB mark. It would be nice to see how much distortion has been introduced when I bring the RMS level up to -16 or -12.
There must be a way of calculating total distortion based on the difference between the current WAV and the same WAV with all > 0 dB peaks truncated to 0 dB. Also maybe there's a way to generate a report of where in the WAV the distortion is worst, and where there is little or none.
Anyway, just a thought I had. For version 6.0?
I noticed that by storing the edited WAV internally in 32-bit floating point format, peaks that exceed the 0 dB threshold (due to volume changes, etc.) are not truncated to 0 dB until you actually save the WAV file to disk.
I do a lot of level equalization from WAVs ripped from CDs, and many older CDs have a low overall volume with some peaks close to the 0 dB mark. It would be nice to see how much distortion has been introduced when I bring the RMS level up to -16 or -12.
There must be a way of calculating total distortion based on the difference between the current WAV and the same WAV with all > 0 dB peaks truncated to 0 dB. Also maybe there's a way to generate a report of where in the WAV the distortion is worst, and where there is little or none.
Anyway, just a thought I had. For version 6.0?